In article <[EMAIL PROTECTED]>, Eli Zaretskii <[EMAIL PROTECTED]> writes:
> I'm not sure this is a bug, though: `D' displays the raw output from > Diff, whose encoding is not clear, because it comes from two different > files that could be each encoded differently. Ediff reads the output > of Diff in raw-text form, precisely for this reason, and that is how > the buffer is displayed to you when you press `D'. What you see is > not gibberish, but the UTF-8 encoding of the non-ASCII characters, > exactly as they are in the written in file1.txt on disk. Right. > I'm not sure we can do any better here, unless we somehow know the > encoding of each of the two files. If efficiency doesn't matter, we can read two files, check which coding systems are used for each file, and parse ourput of diff and decode parts of each file by the corresponding coding systems. But, perhaps, it is good enough to make the command work with universal-coding-system-argument (C-x RET c) (if not yet done), and warn user to use it when the output diff contains eight-bit characters. --- Kenichi Handa [EMAIL PROTECTED] _______________________________________________ emacs-pretest-bug mailing list [email protected] http://lists.gnu.org/mailman/listinfo/emacs-pretest-bug
